The noise generated from vehicular traffic is a major source of environmental pollution. This paper discusses a comprehensive study on the assessment and ANN modeling of noise levels due to vehicular traffic flow at interrupted traffic flow condition, i.e., at SBI (State Bank Intersection) and BSI (Bus Stand Intersection) in Yavatmal city, district place of Vidarbha Region in Maharashtra State (India). Traffic volume data and noise level data were collected simultaneously at ten selected locations. The noise level data were recorded with precision sound level meter (TES-1352 A data logger SLM, IEC651 Type2, ANSI S 1.4 peak hours (morning and evening). Instantaneous noise levels in dBA were recorded (one count per two seconds or at a rate of 30 counts per minute) and processed through Excel by grouped into 15 minutes data to evaluate noise descriptors in the form of Lmax, Lmin, L10, L50, L90, Leq, LNP, TNI ( Traffic Noise Index) & NC (Noise climate).
Artificial Neural Network software (Elite ANN) was utilized to perform modeling. Total traffic, traffic composition (Bus/Truck), LCV (Light Commercial Vehicle), TW (Two Wheelers), bicycle and others in % and carriageway width, distance of sound level meter from pavement edge were considered as input data. The observed input and output data were processed using ANN at interrupted traffic flow conditions. The output was estimated as L10, Leq, LNP, TNI and NC. The performance of the model was tested by root mean square error (RMSE), the mean absolute error (MAE) and correlation coefficient. The model was validated using linear regression analysis where it was observed that there is no significant difference between the observed and predicted output parameters.